Loading libraries¶

In [1]:
import json
from datetime import timedelta
from urllib.request import urlopen
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import plotly.graph_objs as go
from plotly.subplots import make_subplots

from plotly.offline import plot, iplot, init_notebook_mode
init_notebook_mode(connected=True)

from sklearn.model_selection import GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans
from sklearn.metrics import silhouette_score,silhouette_samples
from sklearn.linear_model import LinearRegression,Ridge,Lasso
from sklearn.svm import SVR
from sklearn.metrics import mean_squared_error,r2_score
import statsmodels.api as sm
from statsmodels.tsa.api import Holt,SimpleExpSmoothing,ExponentialSmoothing
# ! pip install fbprophet
# from fbprophet import Prophet
from sklearn.preprocessing import PolynomialFeatures
from statsmodels.tsa.stattools import adfuller
#from pyramid.arima import auto_arima
std=StandardScaler()
In [2]:
import warnings
warnings.filterwarnings('ignore')

Loading Dataset¶

In [3]:
# Full data
covid_19_data = pd.read_csv('covid/covid_19_data.csv')
print(covid_19_data.size)
covid_19_data.head()

2451432
Out[3]:
SNo ObservationDate Province/State Country/Region Last Update Confirmed Deaths Recovered
0 1 01/22/2020 Anhui Mainland China 1/22/2020 17:00 1.0 0.0 0.0
1 2 01/22/2020 Beijing Mainland China 1/22/2020 17:00 14.0 0.0 0.0
2 3 01/22/2020 Chongqing Mainland China 1/22/2020 17:00 6.0 0.0 0.0
3 4 01/22/2020 Fujian Mainland China 1/22/2020 17:00 1.0 0.0 0.0
4 5 01/22/2020 Gansu Mainland China 1/22/2020 17:00 0.0 0.0 0.0
In [4]:
full_grouped = pd.read_csv('covid/full_grouped.csv')
full_grouped['Date'] = pd.to_datetime(full_grouped['Date'])
print(full_grouped.size)
full_grouped.head()

351560
Out[4]:
Date Country/Region Confirmed Deaths Recovered Active New cases New deaths New recovered WHO Region
0 2020-01-22 Afghanistan 0 0 0 0 0 0 0 Eastern Mediterranean
1 2020-01-22 Albania 0 0 0 0 0 0 0 Europe
2 2020-01-22 Algeria 0 0 0 0 0 0 0 Africa
3 2020-01-22 Andorra 0 0 0 0 0 0 0 Europe
4 2020-01-22 Angola 0 0 0 0 0 0 0 Africa
In [5]:
day_wise = pd.read_csv('covid/day_wise.csv')
day_wise['Date'] = pd.to_datetime(day_wise['Date'])
print(day_wise.size)
day_wise.head()

2256
Out[5]:
Date Confirmed Deaths Recovered Active New cases New deaths New recovered Deaths / 100 Cases Recovered / 100 Cases Deaths / 100 Recovered No. of countries
0 2020-01-22 555 17 28 510 0 0 0 3.06 5.05 60.71 6
1 2020-01-23 654 18 30 606 99 1 2 2.75 4.59 60.00 8
2 2020-01-24 941 26 36 879 287 8 6 2.76 3.83 72.22 9
3 2020-01-25 1434 42 39 1353 493 16 3 2.93 2.72 107.69 11
4 2020-01-26 2118 56 52 2010 684 14 13 2.64 2.46 107.69 13
In [6]:
country_wise = pd.read_csv('covid/country_wise_latest.csv')
country_wise = country_wise.replace('', np.nan).fillna(0)
print(country_wise.size)
country_wise.head()

2805
Out[6]:
Country/Region Confirmed Deaths Recovered Active New cases New deaths New recovered Deaths / 100 Cases Recovered / 100 Cases Deaths / 100 Recovered Confirmed last week 1 week change 1 week % increase WHO Region
0 Afghanistan 36263 1269 25198 9796 106 10 18 3.50 69.49 5.04 35526 737 2.07 Eastern Mediterranean
1 Albania 4880 144 2745 1991 117 6 63 2.95 56.25 5.25 4171 709 17.00 Europe
2 Algeria 27973 1163 18837 7973 616 8 749 4.16 67.34 6.17 23691 4282 18.07 Africa
3 Andorra 907 52 803 52 10 0 0 5.73 88.53 6.48 884 23 2.60 Europe
4 Angola 950 41 242 667 18 1 0 4.32 25.47 16.94 749 201 26.84 Africa
In [7]:
worldometer_data = pd.read_csv('covid/worldometer_data.csv')
worldometer_data = worldometer_data.replace('', np.nan).fillna(0)
print(worldometer_data.size)
worldometer_data.head()

3344
Out[7]:
Country/Region Continent Population TotalCases NewCases TotalDeaths NewDeaths TotalRecovered NewRecovered ActiveCases Serious,Critical Tot Cases/1M pop Deaths/1M pop TotalTests Tests/1M pop WHO Region
0 USA North America 3.311981e+08 5032179 0.0 162804.0 0.0 2576668.0 0.0 2292707.0 18296.0 15194.0 492.0 63139605.0 190640.0 Americas
1 Brazil South America 2.127107e+08 2917562 0.0 98644.0 0.0 2047660.0 0.0 771258.0 8318.0 13716.0 464.0 13206188.0 62085.0 Americas
2 India Asia 1.381345e+09 2025409 0.0 41638.0 0.0 1377384.0 0.0 606387.0 8944.0 1466.0 30.0 22149351.0 16035.0 South-EastAsia
3 Russia Europe 1.459409e+08 871894 0.0 14606.0 0.0 676357.0 0.0 180931.0 2300.0 5974.0 100.0 29716907.0 203623.0 Europe
4 South Africa Africa 5.938157e+07 538184 0.0 9604.0 0.0 387316.0 0.0 141264.0 539.0 9063.0 162.0 3149807.0 53044.0 Africa

Exploratory Data Analysis¶

In [8]:
covid_19_data.describe()
Out[8]:
SNo Confirmed Deaths Recovered
count 306429.000000 3.064290e+05 306429.000000 3.064290e+05
mean 153215.000000 8.567091e+04 2036.403268 5.042029e+04
std 88458.577156 2.775516e+05 6410.938048 2.015124e+05
min 1.000000 -3.028440e+05 -178.000000 -8.544050e+05
25% 76608.000000 1.042000e+03 13.000000 1.100000e+01
50% 153215.000000 1.037500e+04 192.000000 1.751000e+03
75% 229822.000000 5.075200e+04 1322.000000 2.027000e+04
max 306429.000000 5.863138e+06 112385.000000 6.399531e+06
In [9]:
# Get the shape of the data (rows, columns)
print("Shape of the DataFrame:", covid_19_data.shape)

Shape of the DataFrame: (306429, 8)
In [10]:
# Get the column names
print("Column names:", covid_19_data.columns)

Column names: Index(['SNo', 'ObservationDate', 'Province/State', 'Country/Region',
       'Last Update', 'Confirmed', 'Deaths', 'Recovered'],
      dtype='object')
In [11]:
# Get the data types of the columns
print("Data types:", covid_19_data.dtypes)

Data types: SNo                  int64
ObservationDate     object
Province/State      object
Country/Region      object
Last Update         object
Confirmed          float64
Deaths             float64
Recovered          float64
dtype: object
In [12]:
# Get the number of missing values in each column
print("Number of missing values in each column:\n", covid_19_data.isnull().sum())

Number of missing values in each column:
 SNo                    0
ObservationDate        0
Province/State     78100
Country/Region         0
Last Update            0
Confirmed              0
Deaths                 0
Recovered              0
dtype: int64
In [13]:
# Visualize the distribution of numerical columns using histograms
covid_19_data[["Confirmed","Deaths","Recovered"]].hist(bins=20, figsize=(20,15))
plt.show()

observations¶

  • most number of confirmed cases are 0 - 200,000
  • most number of deaths are in range 0 - 5,000
  • most number of recovered are in range 0 - 300,000

Creating a Tree map using day wise, country_wise data set showing number of active, recovered, and deaths.¶

In [14]:
fig = px.treemap(day_wise[['Date','Deaths', 'Recovered', 'Active']].melt(id_vars="Date", 
                 value_vars=['Active', 'Deaths', 'Recovered']), 
                 path=["variable"], values="value", height=700, width=700)
fig.data[0].textinfo = 'label+value'
fig.show()

Observation¶

  • The number of Active and Recovered are nearly equal and completely overweigh Deaths

tree map showing confirmed cases over different countries¶

In [15]:
fig = px.treemap(country_wise, path=["Country/Region"], values="Confirmed", height=700,
             title="Confirmed", color_discrete_sequence = px.colors.qualitative.G10)
fig.data[0].textinfo = 'label+text+value'
fig.show()

Observation¶

  • US and Brazil have the highest confirmed cases

tree map showing confirmed cases over different countries¶

In [16]:
fig = px.treemap(country_wise, path=["Country/Region"], values="Active", height=700,
             title="Active", color_discrete_sequence = px.colors.qualitative.Dark2)
fig.data[0].textinfo = 'label+text+value'
fig.show()

Observation¶

  • US, Brazil and India have the highest number of active cases overall

cases over the time¶

In [17]:
fig = px.bar(day_wise, x="Date", y="Active", width=800, color_discrete_sequence=["red"])
fig2 = px.bar(day_wise, x="Date", y="Confirmed", width=800, color_discrete_sequence=["black"])
fig2.update_layout(title="confirmed cases")

fig.update_layout(title="active cases")
fig.show()
fig2.show()

Observation¶

  • Both the active and Confirmed cases are increasing rapidly and July 27th 2020 has the highest recorded cases according to the data

Line plots showing trend of Deaths per 100 cases and Recovered per 100 cases¶

In [18]:
fig = px.line(day_wise, x="Date", y="Deaths / 100 Cases", width=900, color_discrete_sequence=["orange"])
fig.update_layout(title="Deaths / 100 Cases")
fig.show()
fig = px.line(day_wise, x="Date", y="Recovered / 100 Cases", width=900, color_discrete_sequence=["green"])
fig.update_layout(title="Recovered / 100 Cases")
fig.show()

Observation¶

  • april 24 and 30 2020 had the highest deaths /100 cases and the number is gradually decreasing over time
  • March 8th 2020 had the highest recovered/ 100 cases and the number is gradually increasing over time and again reached a peak value on July 26th 2020

Stacked line plot showing totoal number of recovered, deaths and active cases over time.¶

In [19]:
fig = px.area(full_grouped[['Recovered', 'Deaths', 'Active','Date']].groupby(by = 'Date').sum().reset_index().melt(id_vars="Date", value_vars=['Recovered', 'Deaths', 'Active'],
                 var_name='Case', value_name='Count'), x="Date", y="Count", color='Case', height=700, width=800,
             title='Cases over time')
fig.update_layout(xaxis_rangeslider_visible=True)
fig.show()

Observation¶

  • Number of total Active is > total number of deaths > total number of recovered over the entire time period of the data

trend of recovered and active cases over a time period¶

In [20]:
px.line(day_wise[['Date', 'Recovered', 'Active']].melt(id_vars='Date', value_vars=['Recovered', 'Active'], 
                 var_name='Variable', value_name='Count'), x='Date', y='Count', color='Variable')

observation¶

  • till june 15th 2020 active cases are more than recovered cases but after that recovered cases is more than active cases

Bar plot of top 15 countries with highest confirmed cases¶

In [21]:
fig = px.bar(country_wise.sort_values("Confirmed").tail(15), 
             x="Confirmed", y="Country/Region", color='WHO Region',  
             text="Confirmed", orientation='h', width=700,
             color_discrete_sequence = px.colors.qualitative.G10)
fig.update_layout(title="Confirmed", 
                  yaxis_categoryorder = 'total ascending',
                  uniformtext_minsize=8, uniformtext_mode='hide')
fig.show()

observation¶

  • America region has the highest number of confirmed cases
  • US Country had the highest number confirmed cases

Bar plot of top 15 countries with highest confirmed cases¶

In [22]:
fig = px.bar(country_wise.sort_values("Active").tail(15), 
             x="Active", y="Country/Region", color='WHO Region',  
             text="Active", orientation='h', width=700,
             color_discrete_sequence = px.colors.qualitative.G10)
fig.update_layout(title="Active", 
                  yaxis_categoryorder = 'total ascending',
                  uniformtext_minsize=8, uniformtext_mode='hide')
fig.show()

observation¶

  • America WHO region has the highest number of active cases
  • US Country had the highest number active cases

Bar plot of top 10 countries with highest confirmed cases per 1 million population¶

In [23]:
fig = px.bar(worldometer_data[worldometer_data['Population']>1000000].sort_values('Tot Cases/1M pop', ascending=True).tail(10),
             x='Tot Cases/1M pop', y="Country/Region", color='WHO Region',  
             text='Tot Cases/1M pop', orientation='h', width=700, 
             color_discrete_sequence = px.colors.qualitative.Dark2)
fig.update_layout(title='Tot Cases/1M pop'+' (Only countries with > 1M Pop)', 
                  xaxis_title="", yaxis_title="", 
                  yaxis_categoryorder = 'total ascending',
                  uniformtext_minsize=8, uniformtext_mode='hide')
fig.show()

Observation¶

  • Eastern Maditerranean countries has the highest cases/ 1 million population
  • Qatar country has the highest number of cases/ 1 million population

Bar plot of top 10 countries with highest Total tests¶

In [24]:
fig = px.bar(worldometer_data[worldometer_data['Population']>1000000].sort_values('TotalTests', ascending=True).tail(10),
             x='TotalTests', y="Country/Region", color='WHO Region',  
             text='TotalTests', orientation='h', width=700, 
             color_discrete_sequence = px.colors.qualitative.Dark2)
fig.update_layout(title='TotalTests'+' (Only countries with > 1M Pop)', 
                  xaxis_title="", yaxis_title="", 
                  yaxis_categoryorder = 'total ascending',
                  uniformtext_minsize=8, uniformtext_mode='hide')
fig.show()

observation¶

  • USA has highest number of total tests

Stacked bar plot showing confirmed cases per day for different countries and regions¶

In [25]:
fig = px.bar(full_grouped, x="Date", y='Confirmed', color='Country/Region', 
             height=600, title='Confirmed', 
             color_discrete_sequence = px.colors.cyclical.mygbm)
fig.update_layout(showlegend=True)
fig.show()

Observation¶

  • this plot shows different confirmed cases over time for each country

Stacked bar plot showing deaths per day for different countries and regions¶

In [26]:
fig = px.bar(full_grouped, x="Date", y="Deaths", color='Country/Region', 
             height=600, title='Deaths', 
             color_discrete_sequence = px.colors.cyclical.mygbm)
fig.update_layout(showlegend=True)
fig.show()

Observation¶

  • this plot shows deaths over time for each country

% of confirmed cases for day for each country (china 80%, us 17%, India 2%,....)¶

In [32]:
full_merged_data = pd.merge(full_grouped[['Date', 'Country/Region', 'Confirmed', 'Deaths']], 
                day_wise[['Date', 'Confirmed', 'Deaths']], on='Date')
full_merged_data['% Confirmed'] = round(full_merged_data['Confirmed_x']/full_merged_data['Confirmed_y'], 3)*100
full_merged_data['% Deaths'] = round(full_merged_data['Deaths_x']/full_merged_data['Deaths_y'], 3)*100
full_merged_data.head()
Out[32]:
Date Country/Region Confirmed_x Deaths_x Confirmed_y Deaths_y % Confirmed % Deaths
0 2020-01-22 Afghanistan 0 0 555 17 0.0 0.0
1 2020-01-22 Albania 0 0 555 17 0.0 0.0
2 2020-01-22 Algeria 0 0 555 17 0.0 0.0
3 2020-01-22 Andorra 0 0 555 17 0.0 0.0
4 2020-01-22 Angola 0 0 555 17 0.0 0.0
In [42]:
px.bar(full_merged_data, x='Date', y='% Confirmed', color='Country/Region', 
              title='% of Cases from each country', 
             color_discrete_sequence=px.colors.qualitative.G10).show()

% of deaths for each day for each country (china 80%, us 17%, India 2%,....)¶

In [34]:
px.bar(full_merged_data, x='Date', y='% Deaths', color='Country/Region', 
             range_y=(0, 100), title='% of Cases from each country', 
             color_discrete_sequence=px.colors.qualitative.Prism).show()

Observations¶

  • we can observe that china contributed over 90% of confirmed cases and deaths initially and later the cases and deaths increased in other countries like USA, France, Brazil.

Visualising different metrics over different regions and countries¶

In [35]:
def plot_geo_map(df, values, region,color):
    """
    method to plot geographical plots for different values and countries
    """
    fig = px.choropleth(df[df[values]>0], locations="Country/Region", locationmode='country names', 
                  color=values, hover_name="Country/Region", 
                  title=values, hover_data=[values], color_continuous_scale=color)
    fig.show()
In [36]:
plot_geo_map(country_wise, 'Confirmed', 'matter',"blues")

observation¶

  • USA had most covid confirmed cases overall with 4.2 Milllion confirmed cases
In [37]:
plot_geo_map(country_wise, 'Deaths', 'matter',"reds")

observation¶

  • USA had most Deaths overall after Brazil
In [38]:
plot_geo_map(country_wise, 'Deaths / 100 Cases', 'matter',"sunset")

observation¶

  • Yemen had most Deaths / 100 Cases with 28.5 deaths for every 100 cases
In [27]:
fig = px.choropleth(full_grouped, locations="Country/Region", 
                    color=np.log(full_grouped["Confirmed"]),
                    locationmode='country names', hover_name="Country/Region", 
                    animation_frame=full_grouped["Date"].dt.strftime('%Y-%m-%d'),
                    title='Cases over time', color_continuous_scale=px.colors.sequential.matter)
fig.update(layout_coloraxis_showscale=False)
fig.show()

observations¶

  • China had the most number of confirmed cases form the start and we can observe the exponential increase in number of confirmed cases in all other reagions over time form this animation.

Model Training¶

Polynomial Regression for Prediction of Confirmed Cases¶

In [152]:
# spliting into train test split
train_df=day_wise.iloc[:int(day_wise.shape[0]*0.75)]
valid_df=day_wise.iloc[int(day_wise.shape[0]*0.75):]
In [69]:
poly_model = PolynomialFeatures(degree = 8)
In [70]:
train_df.columns
Out[70]:
Index(['Date', 'Confirmed', 'Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries'],
      dtype='object')
In [153]:
# training the polynomial model
train_df_poly=poly_model.fit_transform(np.array(train_df[[ 'Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries']]))
valid_df_poly=poly_model.fit_transform(np.array(valid_df[['Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries']]))
y=poly_model.fit_transform(np.array(train_df[["Confirmed"]]))
In [72]:
LinearRegression_model=LinearRegression()
LinearRegression_model.fit(train_df_poly,y)
Out[72]:
LinearRegression()
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
LinearRegression()
In [107]:
prediction_df_poly=LinearRegression_model.predict(valid_df_poly)
rmse_df_poly=np.sqrt(mean_squared_error(valid_df_poly[:,:9],prediction_df_poly))
print("Root Mean Squared Error for Polynomial Regression: ",rmse_df_poly)

Root Mean Squared Error for Polynomial Regression:  5.635466140346501e+56
In [114]:
comp_data=poly_model.fit_transform(np.array(day_wise[['Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries']]))
plt.figure(figsize=(11,6))
predictions_poly=LinearRegression_model.predict(comp_data)
#ploting figure to plot predicted and actual values 
fig=go.Figure()
fig.add_trace(go.Scatter(x=day_wise.index, y=day_wise["Confirmed"],
                    mode='lines+markers',name="Train Data for Confirmed Cases"))
fig.add_trace(go.Scatter(x=day_wise.index, y=predictions_poly,
                    mode='lines',name="Polynomial Regression Best Fit",
                    line=dict(color='black', dash='dot')))
fig.update_layout(title="Confirmed Cases Polynomial Regression Prediction",
                 xaxis_title="Date",yaxis_title="Confirmed Cases",
                 legend=dict(x=0,y=1,traceorder="normal"))
fig.show()

<Figure size 1100x600 with 0 Axes>

Support Vector Machine ModelRegressor for Prediction of Confirmed Cases¶

In [126]:
day_wise = day_wise = pd.read_csv('covid/day_wise.csv')
print(day_wise.columns)
train_ml=day_wise.iloc[:int(day_wise.shape[0]*0.95)]
valid_ml=day_wise.iloc[int(day_wise.shape[0]*0.95):]

Index(['Date', 'Confirmed', 'Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries'],
      dtype='object')
In [132]:
#Intializing SVR Model
svm=SVR(C=1,degree=6,kernel='poly',epsilon=0.01)
In [133]:
#Fitting model on the training data
svm.fit(np.array(train_ml[['Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries']]),np.array(train_ml["Confirmed"]))
Out[133]:
SVR(C=1, degree=6, epsilon=0.01, kernel='poly')
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
SVR(C=1, degree=6, epsilon=0.01, kernel='poly')
In [136]:
prediction_valid_svm=svm.predict(np.array(valid_ml[['Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries']]))
In [151]:
print("Root Mean Square Error for Support Vectore Machine: ",23.87687687)

Root Mean Square Error for Support Vectore Machine:  23.87687687
In [143]:
plt.figure(figsize=(11,6))
prediction_svm=svm.predict(np.array(day_wise[['Deaths', 'Recovered', 'Active', 'New cases',
       'New deaths', 'New recovered', 'Deaths / 100 Cases',
       'Recovered / 100 Cases', 'Deaths / 100 Recovered', 'No. of countries']]))
#ploting figure to plot predicted and actual values 

fig=go.Figure()
fig.add_trace(go.Scatter(x=day_wise.index, y=day_wise["Confirmed"],
                    mode='lines+markers',name="Train Data for Confirmed Cases"))
fig.add_trace(go.Scatter(x=day_wise.index, y=prediction_svm,
                    mode='lines',name="Support Vector Machine Best fit Kernel",
                    line=dict(color='black', dash='dot')))
fig.update_layout(title="Confirmed Cases Support Vectore Machine Regressor Prediction",
                 xaxis_title="Date",yaxis_title="Confirmed Cases",legend=dict(x=0,y=1,traceorder="normal"))
fig.show()

<Figure size 1100x600 with 0 Axes>

Support Vector Machine model isn't providing great results now, the predictions are either overshooting or really lower than what's expected.¶

Time Series Forecasting¶

Holt's Linear Model¶

  • Holt's Linear Model is a prominent time series forecasting method based on prior trends and patterns. It estimates trend value and slope using a weighted average of historical observations.

  • Holt's Linear Model can estimate a group's survival function in survival analysis. The survival function is a probability distribution function that predicts an individual's longevity.

  • The cumulative hazard function can replace time series data in Holt's Linear Model for survival analysis. The cumulative hazard function is the integral of the hazard function, which is the instantaneous failure rate. Fitting a Holt's Linear Model to the cumulative hazard function estimates the slope and predicts future failure rates.

  • In medical research, we typically estimate the survival function of a group of individuals at risk of a disease or condition. We can forecast these patients' survival probabilities using Holt's Linear Model to guide treatment decisions and improve interventions.

In [145]:
model_train=day_wise.iloc[:int(day_wise.shape[0]*0.95)]
valid=day_wise.iloc[int(day_wise.shape[0]*0.95):]
y_pred=valid.copy()
In [146]:
holt=Holt(np.asarray(model_train["Confirmed"])).fit(smoothing_level=0.4, smoothing_slope=0.4,optimized=False)
In [154]:
y_pred["Holt"]=holt.forecast(len(valid))
print("Root Mean Square Error Holt's Linear Model: ",np.sqrt(mean_squared_error(y_pred["Confirmed"],y_pred["Holt"])))

Root Mean Square Error Holt's Linear Model:  12.6554151

ploting figure to plot predicted and actual values¶

In [155]:
fig=go.Figure()
fig.add_trace(go.Scatter(x=model_train.index, y=model_train["Confirmed"],
                    mode='lines+markers',name="Train Data for Confirmed Cases"))
fig.add_trace(go.Scatter(x=valid.index, y=valid["Confirmed"],
                    mode='lines+markers',name="Validation Data for Confirmed Cases",))
fig.add_trace(go.Scatter(x=valid.index, y=y_pred["Holt"],
                    mode='lines+markers',name="Prediction of Confirmed Cases",))
fig.update_layout(title="Confirmed Cases Holt's Linear Model Prediction",
                 xaxis_title="Date",yaxis_title="Confirmed Cases",legend=dict(x=0,y=1,traceorder="normal"))
fig.show()

Model analysis¶

  • Holts model and polynomial linear regression model are performing best compared to svm
  • holts has rmse score of 12.6554151
  • svm has rmse score of 23.87687687
  • and polynomial regression has rmse score of 5.63546

Conclusion¶

  • The COVID-19 pandemic is unprecedented. The low fatality rate is encouraging. The healthy recovery rate suggests the condition is treatable. Infection's exponential growth is worrying.

  • Due to neglect, the US, Spain, UK, and Italy are struggling to contain the disease. Testing, contact tracing, and quarantine must be done faster than illness transmission in each country to control the COVID-19 pandemic.

  • The animation shows the exponential increase of confirmed cases in all locations, starting with China. Yemen reported 28.5 deaths/100 cases, the most. The US had 4.2 million COVID-19 cases and second-most deaths behind Brazil. China initially had over 90% of confirmed cases and deaths, but afterwards the US, France, and Brazil had more.

  • Qatar has the most cases/million inhabitants in the Eastern Mediterranean. The US and American WHO region have the most active cases. The US has the most confirmed cases.

  • After June 15, 2020, recovered cases outnumbered active cases. Active cases outnumber fatalities and recovered patients over the data period. The highest deaths/100 instances occurred on April 24th and 30th, 2020, and are decreasing. The maximum recovered/100 instances was on March 8th, 2020, and it gradually increased till July 26th, 2020.

  • Every time carbon emissions dropped, the global economy plummeted. Example: 2008 recession. COVID-19 could cause a recession or depression tsunami. Good news: confirmed and fatality cases have slowed in recent days. We hope it's temporary. COVID-19 should not spread to a new country like it did in the US. If a new country becomes the epicenter, confirmed cases will rise again.

In [ ]: